RUN THIS LLM
Search local LLM hardware requirements
Gemma 4 E4B
Google · 8B · Vision
Gemma 4 efficient mid-size model. Multimodal vision-language model with ~8B params delivering strong quality per watt.
VRAM Requirements
| Quantization | VRAM |
|---|---|
| Q4_K_M (smallest) | 4.8 GB |
| Q8_0 (balanced) | 8.8 GB |
| FP16 (full quality) | 16 GB |
Specifications
- Parameters: 8B
- Category: Vision
- Max context: 128K tokens
- System RAM: 12 GB minimum
- HuggingFace: google/gemma-4-E4B-it
Benchmarks
- MMLU: 69 — General knowledge & reasoning (5-shot)
- HumanEval: 52 — Code generation (pass@1)
- MATH: 43 — Competition-level math reasoning
Loading interactive analysis...